Re: [Xen-devel] [PATCH OSSTEST 04/11] TestSupport: introduce set_host_prop

[Roger Pau Monne <roger.pau@xxxxxxxxxx>]

On Tue, Aug 01, 2017 at 02:01:35PM +0100, Ian Jackson wrote:
> Roger Pau Monne writes ("Re: [PATCH OSSTEST 04/11] TestSupport: introduce 
> set_host_prop"):
> > IMHO, I think the right approach is to leave mg-hosts as it is now,
> 
> Yes.
> 
> > and implement a set_property in HostDB/{Executive/Static}.pm and
> > implement a helper in TestSupport that makes use of it
> > ($mhostdb->set_property(...)), do you agree?
> 
> TBH, since this is only being called in the one
> ts-set-host-properties-from-runvars script (or whatever you're calling
> it), I think you can use $mjobdb-> directly.  That's not too bad a
> layer violation.

In the new version that I've sent I've already added a helper to
TestSupport, it's just two lines of code so unless you feel really
annoyed by it I would probably leave it there.

> I think your runvars should probably be named after the ident, not the
> hostname.  That may involve rethinking your encoding, since idents can
> contain _ (hostnames can contain - but not _).

Does it make sense to have the hostname or the ident?

After all ts-set-host-properties-from-runvars is only going to save
properties for the host passed as 'host' ident, and I don't see much
reason for allowing it to support two different idents like src_host
or dst_host, or in general for having a test script that sets host
properties for multiple hosts.

Roger.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel